Topic: Matrix recording steps and FAQ (744T series, and others like it)

[robotaper]

I haven't yet attempted a 4 channel matrix recording but I'm really itching to sink my teeth into my first one.  Has anyone here compiled a step-by-step guide to matrix recording using the 744 for soundboard/2 mic arrangements?  I'm looking for recommendations in the areas of microphone configurations, gain levels, file types (mono .wav vs. poly .wav), distance to sound source to delay metering, and recommended post production methods for maximizing the 744's capabilities.

From what I've seen here most are recording their matrix poly .wav, and then combine the 4 tracks in post production.  This is where my technical prowess begins to wane a bit.  Any help or suggestion would be appreciated.

If this is not something that has been compiled I will be happy to do so.

Thanks.

[Fatah Ruark (aka MIKE B)]

With 4 channels you don't have to worry about delay.

From what I've gathered a matrix with both sources using the same word clock is pretty easy. Just record your 4 channels, then plug them into wav editting software (like Wavelab) and adjust the delay so both sources match up. Then you can play with levels to make the recording sound like you'd like it to. Once you've got it dialed render the 4 channels down to 2 and you should be good to go unless you want to add some compression, eq, etc to the mixed source.

In Wavelab you'd want to mix the channels with the Audio Montage option (under File on the menu).

[su6oxone]

Matrixing is easy with Vegas.  Just align the tracks and stretch one to match the other.  It can be done in 5-10 minutes. 

[Brian Skalinder]

With 4 channels you don't have to worry about delay.

Not necessarily.  Delay is a function of time-arrival differences between multiple recorded sources and not really contingent on the number of channels.  For example, a 2-ch (or 3- or 6-ch for that matter) AUD source may have a delay relative to the SBD source, depending on distance and any board/stage routing.  Depending mainly on the distance (and therefore the length of delay), one might need to align the AUD source to the SBD source, or vice versa.

If one uses the same clock for all sources (whether using a single ADC in a 4-ch recorder like the 744, or linking clocks between multiple recorders), one needn't worry about drift...which is a much bigger PITA.

[stevetoney]

With 4 channels you don't have to worry about delay  DRIFT .

I'm sure he probably knows what you mean, but the issue of drift is eliminated when you have a 4 channel recorder, not delay.  

Drift is where the two recordings, when recorded on different machines, start at the same time but become gradually farther apart from each other because one recorders speed was different from the other.  As the sources are played, they drift farther apart the longer the recording goes.

As you point out, delay is still possible in a 4 channel recording because there can potentially be a time difference between, say the SBD tracks and the AUD tracks.  This is caused by the short amount of time it takes for sound to travel through the air to reach the audience mics, whereas that time delay doesn't happen on the SBD tracks since the sound goes straight into the stage mics then electronically to the SBD mixer.  

[datbrad]

With 4 channels you don't have to worry about delay.

Not necessarily.  Delay is a function of time-arrival differences between multiple recorded sources and not really contingent on the number of channels.  For example, a 2-ch (or 3- or 6-ch for that matter) AUD source may have a delay relative to the SBD source, depending on distance and any board/stage routing.  Depending mainly on the distance (and therefore the length of delay), one might need to align the AUD source to the SBD source, or vice versa.

If one uses the same clock for all sources (whether using a single ADC in a 4-ch recorder like the 744, or linking clocks between multiple recorders), one needn't worry about drift...which is a much bigger PITA.

I was thinking that he meant not worrying about setting the delay live on the 744T, because all 4 channels are being recorded separately, and any time alignment can be done in post. In the old days, matrix recordings were done live on the fly, and so if the SBD delay was not calculated correctly, the resulting 2 channel live mix down would not be aligned correctly, and there was nothing you could do about it later. Just my take.

[Brian Skalinder]

I was thinking that he meant not worrying about setting the delay live on the 744T, because all 4 channels are being recorded separately, and any time alignment can be done in post.

Ah, of course...reading it again, that makes complete sense.  Thanks, DATBRAD.

[datbrad]

I was thinking that he meant not worrying about setting the delay live on the 744T, because all 4 channels are being recorded separately, and any time alignment can be done in post.

Ah, of course...reading it again, that makes complete sense.  Thanks, DATBRAD.

Just glad to know I was able to contribute something useful, since I don't do 4 channel myself.

[robotaper]

So it seems for the most part that all touch up work (matching of the two sources, adjustment of volume to achieve desired sound quality, etc.) is done with a capable post production PC program.  Wave Lab seems to be the most common application used, although it carries a pretty hefty price tag for a new copy.  Any suggestions on acquiring a discounted copy of this App or if you're aware of an alternative program that achieves similar results?

[SmokinJoe]

The delay between AUD and SBD is generally about ".001 seconds per foot" from the stacks to the mics, so you will need to trim something off the start of the AUD tracks to align.  I generally align by eye and ear, but this is a good approximation.  If my mics are 20' from the stacks, the amount that I need to trim off the front of the AUD waveform is about .020 seconds.  It might be a bit more... like .025 or so, but it's not .100 or .200.

I generally pick a point in the tracks like a point where it goes from flat line to a spike, figuring out how much I need to chop, then repeat the check at few more points, and kind of do a mental average... using that .001/foot as a sanity check. 

I've heard different people say:
a) align for the start of an impulse
b) align for a peak
I've done both, and frequently average between the two.

Most of my mixes are about 50/50 SBD/AUD, unless the SBD is really, really good.  In this case I would Amplify both so the peaks hit about -2db.  If they both go to 0db, then when they combine to +3db the clipping can become very audible.

This is not an exact science... at least not to me yet.

I use Audacity... it runs on both my Linux and Windows box, and it's free.

[capnhook]

I concur with all that's been said here.  I'd add that I've also had the occasion to make live matrix recordings through a four channel mixer to put the two sources together (SBD and AUD).  Under a dancetent setting, where there are all sorts of time delays and reflections coming together in an appealing way, I haven't adjusted at all for delay.......I try to place the mics closer to the band than at the optimal equilateral distance, and then I use a good closed-ear set of headphones to balance the mix, searching for that rich natural reverberation that is so hard to get any other way.   

[stevetoney]

I concur with all that's been said here.  I'd add that I've also had the occasion to make live matrix recordings through a four channel mixer to put the two sources together (SBD and AUD).  Under a dancetent setting, where there are all sorts of time delays and reflections coming together in an appealing way, I haven't adjusted at all for delay.......I try to place the mics closer to the band than at the optimal equilateral distance, and then I use a good closed-ear set of headphones to balance the mix, searching for that rich natural reverberation that is so hard to get any other way.   

Nice points capnhook.  We have a venue here in Pittsburgh that attracts quite a few acts that we (team pittsburgh) like to record.  The venue is taper friendly, but they make us set our rigs up against a drink rail about half way back in the venue.  The sound breaks up a little by the time you get back to that spot, but if you walk a mere 3 or 4 paces forward, the sound is usually very direct and nice.  What seems to happen therefore is that the recordings can be hit and miss at this place because some nights the mix gets the direct sound back to our location and some nights it breaks up just in front of us.  The point being, it would be really nice to be able to dial it in like you said to get just the right mix of direct to reverb mixture every night.

[Brian Skalinder]

Well, I don't agree 100% that you don't have to disregard the effect of delay in a multi track recorder.

I don't think anyone is saying to disregard delay outright.  I think the point most are making is to not worry about delay during recording, but to address it during post-production.

[stevetoney]

I do know that when I last pulled a live mix aud/sbd with my UA5 that there was a slight but noticeable "echo/verb" quality to the stereo wave recorded.  I can't imagine the 744 acting any different.

Right.  Of course the echo is caused by the two sources, when played simultaneously, being slightly out of synch with each other by some milliseconds.  The farther the AUD mics were from the stage, the more milliseconds out of synch they are (and the more the reverb effect you'll hear). 

For this delay effect, it's irrelevant whether you're using a 744, an R-44 or any other time synched recording setup (two time synched 702's, for example)...the delay exists because of the laws of physics.  The sources will still be out of synch because of the delay...and this is fixed by sliding one of the sources forward or backwards (relative to the other source) by the same number of milliseconds that the two sources are out of synch.  Once the two sources line up exactly with each other, you hear no more echo/reverb effect when the two sources are played simultaneously.

I think what some are saying is that the 744 has electronics capabilities that allows for internal compensations to be made (to add those milliseconds of delay to one of the sources, for example) during capture so that the two sources will be synched as the recording is made. 

[baustin]

drift = don't need to worry about
delay = you do need to worry about. factors such as cable lengths, mic placement (onstage vs room), recorder location (i occasionally tape on the stage so band mics go to the board then back to the stage via the in house snake), etc

i'd recommend 4 mono wav files with 1GB file splits. wavelab 5 and below has a 2GB file limit. when you matrix, you'll take 4GB of files and drop it down to 2GB. if i'm doing a sbd + mic matrix, i like for the levels to peak around -8db. that way i can add gain in post rather than drop levels.

if running 4 mono files you can also begin to think outside of the box. maybe 2 onstage mics, 1 room mic and 1 sbd channel (many smaller venues run mono so you can apply the sbd feed to the L/R channels evenly in post)